Dasaratha, the founder of the three Nâgârjuni Caves, had actually been preserved down to a comparatively late period, and was then ignorantly referred by the Brahmans to the early king of Angga, instead of to the Maurya Prince of Magatha. Regarding the name of Sudâma or Sudhâma, I am unable to offer any conjecture; but Viswamitra was one of the most celebrated of the seven Rishis, or great Brahmanical Saints.

The silence of Hwen Thsang regarding the caves has been already noticed; but I have a suspicion that he had heard of the celebrated spring of the Pátâl Gangâ at the foot of the Barâbar Hill. According to his account, there was a famous spring of pure water situated at 30 li (or 5 miles) to the north of Gaya.* Now, as I could not hear of any spring to the northward of Gaya nearer than Barâbar, I would suggest that Hwen Thsang's distance of 30 li should be corrected to 130 li (or 21 miles), which would make his famous spring agree exactly with the position of the Pátál Ganga, according to the distance by road, which is 13 miles to the Bela Dâk Bungalow + 6 to the Kauwa-Dol Hill + 2 more to the Pâtâl Gangâ. Hwen Thsang adds that "the Indians, following an ancient tradition, called this spring the holy water' (l'eau sainte), and that at all times whoever drank of it, or bathed in it, was instantly purified from the stain of his sins. Now the source of the Pâtâl Gangâ is still held in such esteem that, according to Buchanan, from 20,000 to 50,000 people assemble annually in the middle of the month of Bhadrapada to bathe in its waters, and about 500 people bathe daily during the whole of that month.

Should this identification be correct, it would seem to be almost certain that towards the middle of the seventh century of our era, not only were these caves occupied by the Brahmans, but the very memory of their Buddhist origin had either been forgotten or was carefully concealed.

XIV. DHARAWAT.

The Dharawat group of hills lies immediately to the northward of the Barâbar hills, about 1 mile distant. There are two distinct ridges running from west to east, that to the

* Julien's Hwen Thsang, II., 455.

south being nearly two miles in length with three peaks named Saleya, Gureya, and Dhaoli.* The nearest road from Barabar to Dharâwat lies through a pass between the Gureya and Dhaoli Hills. The northern ridge consists of a single hill named Ratani, which in former days was occupied by some establishment of the Buddhists. On the northern slope of the hill there are two brick terraces which have been built up against the rock. The eastern terrace is 60 feet long by 20 feet broad, and 50 feet above the plain. Near the top the solid brick-work can still be seen for 20 feet in height, below which the brick rubbish reaches to the foot of the hill. The second terrace lies more than 200 feet to the westward of the other; it has a front of 250 feet, but its height is not more than 15 feet above the plain. On this terrace there are two broken Buddhist figures, and beneath it there are four others, of which one bears the usual Buddhist formula of "Ye Dharmma hetu prabhava, &c.," in characters of the 9th or 10th century.

To the north of the Ratani Hill there is a large tank called Chándokhar Tâl, 2,000 feet in length and 800 feet in width. On the eastern embankment there is a new temple to Mahâdeva, only three years old, and close beside it a very small old temple to Narsingh. Outside this temple there is a very fine life-size statue named Bhairav. The figure stands under a thick stem of lotus which forms an arch overhead, and from which little curling branches strike off on both sides, ending in lotus flowers which support tiny figures of men, women, and animals. The statue has twelve arms, and bears in the head-dress a small figure of Buddha squatted with hands in lap. I recognized it at once as a statue of the famous Bodhisatwa Avalokiteswara. Beside the statue, there are several sculptured stones containing rows of Buddhas, and also several fragments of votive stupas, and two slabs with representations of the Nava-graha, or "nine planets." There are also numerous fragments of sculpture under a Pipal tree close by, two of which bear inscriptions in characters of the 9th or 10th century.

To the north-east of the Chândokhar Tâl there is an extensive mound of brick ruin, which is probably only the remains of the former town of Dharâwat. In the north-west

See Plate No. XVIII.

corner of this mound there are two small eminences, which may be the remains of temples, but as the surface of the mound now presents nothing but small fragments of bricks, all the larger bricks having been removed to furnish materials for the present village, it is quite impossible to say what kind of buildings may once have stood upon it. All that can be inferred, I think, from the present remains is, that Dharâwat must at one time, probably about the 8th or 9th century, have been the seat of a considerable Buddhist community. Major Kittoe paid a hurried visit to Dharâwat by moon-light. He notices the twelve-armed figure, which he calls a Buddhist sculpture, as being very remarkable.

XV. BESARH.

The village of Besárh, or Besádh in Nagari characters, is situated 27 miles, a little to the east of north from Patna, and 20 miles from Hâjipur on the left bank of the Ganges. Both the distance and direction from Patna point to this place as the representative of the ancient Vaisáli. The name also is the same, as it is written Besárh by Abul Fazl in his Ain Akbari.* Now, Hwen Thsang places the King's Palace in Vaisâli at 120 li, or 20 miles, to the east of north from the northern bank of the Ganges opposite Pâtaliputra, that is, from the present Hajipur. He also describes the King's Palace as being from 4 to 5 li (from 3,500 to 4,400 feet) in circuit, which agrees with the size of the ruined fort now called Raja Bisál-ka-garh, which is 1,580 feet long and 750 feet broad inside, or 4,660 feet in circuit round the crest of the mound. This almost perfect coincidence of name, position, and dimensions, seems quite sufficient to place the identification of Besârh with Vaisâli beyond all reasonable doubt. I will, therefore, now proceed to describe the objects of interest that still remain in Besârh and the neighbouring village of Bakhra, which will afford further proof of the identity of Besârh and Vaisâli.

These ruins were visited by Mr. J. Stephenson in 1834, and described by him in Prinsep's Journal. They consist of two distinct groups, one at Besârh itself, and the other

* Gladwin's Translation, II., 198.

+Julien's Hwen Thsang, II., 399. To Swetapura 90 li, plus 30 li to the Ganges. In Vol. I., p. 137, the distance to Swetapura is stated to be 100 li.

Bengal Asiatic Society's Journal, 1835, p. 128.